ARTICLE pubs.acs.org/JPCB
Kinetics Inside the Protein: Shape of the Geminate Kinetics in Myoglobin Djemel Hamdane,† Laurent Kiger, Gaston Hui-Bon-Hoa, and Michael C. Marden* Inserm U779, University Paris 11, 94275 Le Kremlin-Bic^etre, France ABSTRACT: Synchronized kinetics of ligand binding to a buried active site offers a look inside the protein. Photodissociated ligands are initially alongside their original binding site, so the recombination kinetics describes the trajectory for direct (geminate) rebinding or escape from the protein for the subsequent (bimolecular) rebinding phase. In the model case of myoglobin in water, most of the ligands escape; to better observe the geminate phase, high viscosity cosolvents were used: the kinetics were characterized by multiple barriers and a distribution of rates. An alternative method to enhance the fraction of geminate phase is the application of high pressure which closes the ligand migration channel; in this case of low viscosity without cosolvents, the geminate phase is closer to a simple exponential behavior. Samples with glycerol display the extended geminate kinetics, while samples in water under pressure do not.
’ INTRODUCTION Proteins such as Hb and Mb are models of structure function studies, yet they represent some of the most difficult cases for such correlations. They are not lock and key systems, as they function without protein partners. The active site for oxygen binding is not at the surface, and the crystallographic structure does not reveal a ligand binding pathway. The protein thus appears as a closed black box, and protein fluctuations are required to allow the passage of ligands. Furthermore, studies of hundreds of species and mutants with similar 3D structure may display large variations in ligand affinity. At the other extreme, various species may have a modest primary sequence homology and yet maintain quite similar functional properties. In these cases, the structure function relationship requires a higher degree of resolution. Much of the knowledge of the internal protein mechanism was revealed through kinetics studies. Methods such as flash photolysis allow a probe of the time dependence of the intramolecular reactions,1,2 while sacrificing the structural information. Only recently has time-resolved crystallography begun to directly probe the ligand migration through various molecular gates and pockets.35 The initial studies with cosolvent showed a complicated recombination curve represented by a series of barriers between the exterior of the protein to the internal binding site.1 As the temperature was lowered (higher viscosity) the ligand escape was hindered and one probed the innermost barriers.2,6 At very low temperatures, the features of the innermost (chemical) barrier could be isolated; the overall shape of the kinetics was a power law (population vs time as a straight line on a loglog plot) and could be simulated by a distribution of rate coefficients.1 In high viscosity solvents where the multiple barriers were evident, the slope near 1/2 in many cases suggested a diffusional like mechanism. However, the pure diffusion model was a bit too r 2011 American Chemical Society
smooth relative to the series of phases. A simple series of barriers was a bit too jagged (stairway effect); the overall model thus involved a series of barriers, coupled with a distribution of rates that were observed at very low temperatures. In cases where the geminate rebinding was faster, such as for ligands O2 and especially NO, cosolvents were not needed, and the geminate phase was less extended. Certain mutants also displayed a different shape for the geminate phase.7,8 One could demonstrate a strong dependence of the ligand migration inside the protein on the solvent viscosity,6 but it was impossible to determine whether the solvent had an additional effect. Since the external viscosity effected the internal rate parameters (except for the innermost chemical barrier), it was clear that there is a strong coupling between the solvent and protein viscoelastic properties. The protein is not an inert sphere in the solvent. The main goal of this study is to reconsider the number of barriers involved in the ligand binding process. High pressure techniques offer an alternative method of enhancing the geminate fraction.9,10 Under these conditions, the solvent remains in a low viscosity state.11,12 Moderate pressures of samples with cosolvents did not change the overall form of the kinetics.13 At sufficiently high pressures, one observes mainly the geminate phase and can thus determine the effect of various cosolvents.
’ EXPERIMENTAL PROCEDURES Protein Preparation . Lyophilized myoglobin from horse heart was purchased from Sigma and purified on a G25 column. A further purification was made with an FPLC Akta Purifier Received: July 30, 2010 Revised: February 24, 2011 Published: March 23, 2011 3919
dx.doi.org/10.1021/jp107168b | J. Phys. Chem. B 2011, 115, 3919–3923
The Journal of Physical Chemistry B (Amersham Pharmacia Biotech) using a column Hitrap DEAE sepharose Fast Flow. The High-Pressure Apparatus made of maraging steel is capable of measurements to 700 MPa (7 kbar). The high pressure cell with sapphire windows was installed in the sample compartments of the Cary 3E or SLM 8000 spectro-fluorimeter. The inner sample holder is a 5 5 mm quartz cuvette. The absorption spectra (between 280 and 700 nm) were measured as a function of hydrostatic pressure up to 600 MPa. The spectral results were corrected for the solvent compressibility,11 which (for water) is initially 4% per kbar and a total of 15% at 600 MPa. One also needs to take into account the shift versus P of the Soret band, especially for the CO species, in order to determine the fraction dissociated (based on the observed signal amplitude). All samples were in 50 mM Tris-HCl buffer at pH 7, since there is a weaker pH variation versus pressure (0.02 units per kbar) compared to 0.33 units for potassium phosphate buffer.14 Flash Photolysis at High Pressures. The high pressure cell was built into an SLM 8000 fluorometer for use of the light source and excitation monochromator. The photolysis is achieved by a 10 ns YAG laser pulse at 532 nm (Quantel, France) at a right angle to the transmission beam, detected by a photomultiplier and recorded on a Lecroy (Waverunner LT342L) oscilloscope.9 A large (6 mm) diameter fiber optic light guide was used to bring the photolysis beam to the sample cell. This results in some loss of intensity but provides a more uniform beam. For photolysis yield measurements, one does not want an excess of energy since certain hemes may be photolyzed more than once; the photolysis was approximately 50% of the hemes. The detection wavelength was varied from 410 to 440 nm in steps of 2 nm, for light of bandwidth 2 nm. The observed kinetic amplitudes were then normalized based on the static spectra to obtain the fraction of hemes with ligand. Geminate vs Bimolecular Kinetics . After photodissociation a certain fraction of the dissociated ligands rebind to the same iron atom without leaving the protein; this geminate phase occurs typically on the ns time scale under normal conditions. The fraction of ligands that escape from the protein will compete with other solvent ligands for binding sites; this bimolecular phase is generally on the μs to s time scale. To describe the overall kinetics, we consider a simplified A - B S reaction scheme, where state B is the correlated (or geminate) pair formed by the photodissociation. Since photodissociation places the ligand in state B, the bimolecular yield will depend on the relative rates for rebinding (B to A) or escape to the solvent (B to S). The inner barrier (BA) corresponds to the “chemical” binding step to the iron atom, and the rate kBA generally increases with T but is nearly independent of solvent viscosity.6 The intrinsic rate coefficient for the ATBTS model can be extracted from the four observable parameters: the geminate rate λ gem = k BA þk BS, the bimolecular or “on” rate λ on = k SB *k BA /(k BA þk BS) = k SB *(1- f bimol) when k AB is negligible, the overall dissociation or “off” rate λoff = k AB * k BS/(k BA þ k BS ) = k AB *(fbimol), and the fraction bimolecular after photolysis f bimolecular = k BS/(kBA þ k BS). Simplified forms are obtained when the overall association rate is barrier limited (k BA , k BS corresponding to a high bimolecular yield) or migration limited (k BA. k BS resulting in a low fraction bimolecular). From the classic thermodynamic relation, k = amplitude exp[-P(ΔVI)/RT], the change in the rate versus pressure allows an estimation of a difference in volume. For a rate coefficient, the
ARTICLE
Figure 1. CO recombination kinetics after photodissociation. CO recombination to horse heart Mb in 80% w/w glycerol (lower panel) or water (upper panel) at pH 7 in Tris buffer, 15 °C, at the pressure (MPa) indicated. Increasing pressure closes the ligand migration channels and results in less bimolecular phase. The samples in glycerol show the extended geminate phase, which covers several orders of magnitude in time, while samples in water display a nearly exponential shape for the interior (μs) reaction. The final (bimolecular) phase in both cases has the shape of a simple exponential.
activation volume ΔVI corresponds to the difference between the top of the barrier (transition state) to the initial state.
’ RESULTS AND DISCUSSION The kinetics, with and without cosolvent, are shown in Figure 1. In both cases, higher pressure results in less bimolecular phase. With the cosolvent of 80% w/w glycerol, the multiple phases are present, as previously observed for samples in a high viscosity environment. For the sample without cosolvent, the viscosity remains low;11 in this water like environment one can observe a nearly exponential geminate phase. For samples with cosolvent, simulation of the kinetic curves requires both multiple barriers (A-B-C-D-S, lower section of Figure 2) as well as a distribution of rates for the inner barrier.1 This has been the reference model for many years; note that the barrier B-A in this model is considered as a distribution of rates.1 For the protein in water at high pressure, a simpler A-B-S model 3920
dx.doi.org/10.1021/jp107168b |J. Phys. Chem. B 2011, 115, 3919–3923
The Journal of Physical Chemistry B
Figure 2. Multiple barrier scheme (energy vs reaction coordinate) for ligand binding to Mb. The lower pane displays the classical model of multiple barriers,1 where high viscosity solvents (lower panel, dotted line) slow all the outer barriers,6 but not the inner, chemical binding (BA) step. High pressure applied to an aqueous solvent (upper panel) also closes the protein migration channel in water and reveals a simpler binding scheme; note that, unlike the high viscosity effect, high pressure changes the ligand affinity.
is sufficient (top section, Figure 2) with the dissociated ligand being only one step from the solvent. We can consider four cases of the solvent conditions. 1) Water at ambient T and P (top curve in Figure 1, upper section). There is generally little geminate phase, and the geminate kinetics can therefore be simulated with a single15 or multiple barriers. 2) With cosolvent (Figure 2, lower section): high viscosity slows migration through the protein and thus decreases the fraction of bimolecular recombination. 3) High pressure (Figure 2, upper section) also slows migration of the ligand and accelerates the rate of geminate rebinding. 4) High pressure and cosolvent (data of Figure 1, lower section) also display the extended geminate phase, indicating a correlation of this effect with the presence of cosolvent. The amplitudes and rate coefficients versus pressure are shown in Figure 3. With or without cosolvent, there was a reversible variation of the fraction geminate phase. The shift in the barriers is globally described by the dotted lines in Figure 2. In both cases the outer barriers increase, reflecting the slower migration through
ARTICLE
Figure 3. Amplitude (top section) and rate coefficients versus pressure. At higher pressures, the rates of entry and escape from the protein (B to S in Figure 2) are greatly reduced, effectively closing the protein and leading to a decrease in the fraction of bimolecular phase (top panel); the solid line is a fit with a single activation volume for each rate coefficient. The observed rates λ (lower panel), for the geminate and bimolecular (on) rates are a combination of the individual rate coefficients k (A-B-S).
the protein. High viscosity does not effect the inner (B to A) chemical barrier; the outer barriers are changed in a symmetrical fashion which does not change the ligand affinity. This is not the case for high pressure, which leads to a higher ligand affinity essentially through a lower ligand dissociation rate;9 this is counterintuitive since one would expect high pressure to favor the smaller volume (without ligand) state. Apparently the protein is designed to be more stable with the ligand, and the cavity of the deoxy state costs more energetically (see activation volumes in Figure 4). Spectral Analysis . An additional effect was observed at very high pressures. Above 500 MPa for MbCO, the overall amplitude of the kinetics began to decrease. Even after correction for the difference spectra of the CO and deoxy forms versus pressure (Figure 5), there was a drop of about 10% by 600 MPa. This could indicate a drop in the quantum yield for the photodissociation 3921
dx.doi.org/10.1021/jp107168b |J. Phys. Chem. B 2011, 115, 3919–3923
The Journal of Physical Chemistry B
Figure 4. Activation volumes for the two barrier, three state (A-B-S) scheme of ligand binding to Mb (upper panel of Figure 2). The outer barrier is highly sensitive to pressure and slows ligand migration in both directions, while the inner chemical binding (BA) has a smaller pressure effect and favors ligand binding.
Figure 5. Spectral shift of MbCO absorption versus pressure. The difference spectra are the absorbance under pressure, measured every 50 MPa, minus the spectrum at 1 bar. The spectra were also measured for deoxy Mb (not shown) in order to estimate the fraction photodissociation versus pressure for the photolysis experiments (Figure 1).
process, or faster phase, beyond the resolution of our 10 ns system; if so, it would be much faster than any rate previously reported for CO binding to Mb. One could also argue that the spectrum of the “B” state (deoxy Fe, but with ligand still inside the protein) is different. As mentioned above, there is no open channel for ligand migration; the protein is thus in a stressed state and sufficiently high pressures would further stress the system. The absorption spectrum for the B state could then differ from that for “S”; if this were the case, one would expect to recover the spectral difference when the ligand finally escapes, but this was not detected. Several additional effects could influence the observed amplitude by a few percent. One might expect a small loss of protein, but we
ARTICLE
have previously shown by fluorescence quenching studies that Mb is stable in this pressure range.16 High pressure also favors His binding to the iron, forming a His-Fe-His hexacoordination as observed for neuroglobin9,17 and Hb alpha chains;18 this effect can be seen in the spectrum of deoxy Mb at pressures above 500 MPa. Note also that the photodissociation process is not in saturation, a necessary condition for comparing the total flash signal vs pressure; about 50% of the hemes are photodissociated, so slight changes of the optical conditions could also influence the fraction dissociated. These additional effects could explain the decrease in the amplitude of the flash signal at very high pressures. The main new result is an apparent collapse of the multi barrier model. At high pressure the geminate rebinding occurs mainly as a single exponential phase. Other proteins were also studied. While their geminate phase was too rapid for a detailed analysis, there was no extended geminate phase at longer time scales. This suggests that in all cases there is no extended geminate phase (wide distribution of rates) under aqueous conditions. Cosolvent Effect . Addition of a few % glycerol (concentrations on the order of 1 M) was sufficient to provoke an observable difference in the kinetics (data not shown). This indicates that the solvent effect occurs at a low % cosolvent, rather than for significant changes in viscosity. This would allow a clear separation of the three effects: pressure, viscosity, and cosolvent. There is definitely an interaction of the cosolvent with the protein. High solvent viscosity slows the migration reaction rates inside the protein indicating a high coupling between the solvent and protein dynamic properties. But it is not clear whether there is an additional solvent effect, due to the binding of solvent molecules which may enter the ligand binding channel and thus perturb the binding pathway. If the main channel is blocked by glycerol molecules, the photodissociated ligand may be forced to probe alternative pathways. This could lead to the multitude of observed barriers. Note that from the kinetics alone, it is difficult to distinguish a series of barriers from several parallel pathways. The cosolvent may simply modify the rates for the transitions of the main pathway, perhaps allowing a better look at certain steps, or the cosolvent may lead to totally new pathways. Similar results were obtained with other globins, such as the alpha chains of hemoglobin18 and neuroglobin.9 In all cases, samples with glycerol displayed the extended geminate phase over several orders of magnitude, with geminate recombination extending to the ms time scale at sufficiently low amplitudes of the bimolecular phase. For the same proteins in water, higher pressure shows a drop in the fraction of bimolecular phase but without the extended geminate phase; the geminate kinetic curves are flat after 100 ns. Observable States . The high pressure data suggest that a simpler model involving a single intermediate state (A-B-Solvent) is sufficient for describing the ligand pathway. This does not necessarily mean that the other barriers are not present. When dealing with perturbation methods new states may become observable, since the external effects may lift a degeneracy allowing a better look at the detailed mechanism. One of the goals of these methods is to determine all the states and rates, even if they are not the dominant forms physiologically. There is thus a question as to whether the solvent is simply modifying existing rates or creating new barriers. If cosolvent molecules are blocking the dominant pathway, then the ligand may be forced to find new ones (in parallel or in series), which could reflect the wide variation in rebinding rates; if this is the 3922
dx.doi.org/10.1021/jp107168b |J. Phys. Chem. B 2011, 115, 3919–3923
The Journal of Physical Chemistry B case, one would be observing possible states, but they may not play a central role in terms of the main physiological mechanism. From Figure 2 (lower panel), one could imagine a transition between the two models. If the external barriers (AB and DS) are higher than the internal steps (BC and CD), then the wells B-CD could behave as one large reservoir. In fact the time coefficient on the order of 1 μs observed under aqueous high pressure conditions is much slower than the previous values for the step BA and might be better associated with a binding of C to A in the multiple barrier scheme. Length vs Escape Time . A simple analysis of the escape time indicates an extremely long time to travel such a short distance. Independent of the models and experimental conditions, times on the order of μs are required for the ligand to travel at most 1 nm. Simple diffusion in water would require about 1 ns to cover the same distance by a random walk. The protein interior thus behaves as if it were 1000 times more viscous than water. A multiple barrier model involving a slow migration was thus reasonable. The high pressure results suggest a single intermediate chamber. This could of course be a series of pockets that have a rapid equilibrium relative to two main outer barriers, as suggested from some simulations.19,20
’ CONCLUSIONS High pressure slows the passage through protein channels and thus enhances the fraction of geminate recombination after ligand photodissociation. The observed rate is slower that previously reported for the innermost barrier. This technique thus enables observation of the geminate kinetics under conditions of low viscosity without cosolvents, thus affording a direct measure of the solvent effect. As for previous studies, samples with glycerol show the extended geminate phase covering several orders of magnitude in the rate distribution. For the present study with samples without cosolvents, a simpler scheme is sufficient, where the photodissociated ligand is only one barrier away from the solvent. Under water-like conditions, the geminate phase takes on a more exponential behavior. Only the general features of the new kinetics are present here; future detailed molecular models should take into account this new data set. ’ AUTHOR INFORMATION
ARTICLE
(2) Fenimore, P. W.; Frauenfelder, H.; McMahon, B. H.; Parak, F. G. Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 16047–16051. (3) Srajer, V.; Ren, Z.; Teng, T. Y.; Schmidt, M.; Ursby, T.; Bourgeois, D.; Pradervand, C.; Schildkamp, W.; Wulff, M.; Moffat, K. Biochemistry 2001, 40, 13802–13815. (4) Schmidt, M.; Nienhaus, K.; Pahl, R.; Krasselt, A.; Anderson, S.; Parak, F.; Nienhaus, G. U.; Srajer, V. Proc. Natl. Acad. Sci. U.S.A. 2005, 102, 11704–11709. (5) Bourgeois, D.; Vallone, B.; Arcovito, A.; Sciara, G.; Schotte, F.; Anfinrud, P. A.; Brunori, M. Proc. Natl. Acad. Sci. U.S.A. 2006, 103, 4924–4929. (6) Beece, D.; Eisenstein, L.; Frauenfelder, H.; Good, D.; Marden, M. C.; Reinisch, L.; Reynolds, A. H.; Sorensen, L. B.; Yue, K. T. Biochemistry 1980, 19, 5147–5157. (7) Petrich, J. W.; Lambry, J. C.; Kuczera, K.; Karplus, M.; Poyart, C.; Martin, J. L. Biochemistry 1991, 30, 3975–3987. (8) Dantsker, D.; Roche, C.; Samuni, U.; Blouin, G.; Olson, J. S.; Friedman, J. M. J. Biol. Chem. 2005, 280, 38740–38755. (9) Hamdane, D.; Kiger, L.; Hui Bon Hoa, G.; Dewilde, S.; Uzan, J.; Burmester, T.; Hankeln, T.; Moens, L.; Marden, M. C. J. Biol. Chem. 2005, 280, 36809–36814. (10) Brindell, M.; Stawoska, I.; Orzez, L.; Labuz, P.; Stochel, G.; van Eldik, R. Biochim. Biophys. Acta 2008, 1784, 1481–1492. (11) Vedam, R.; Holton, G. J. Acoust. Soc. Am. 1968, 43, 108–116. (12) Bett, K. E.; Cappi, J. B. Nature 1965, 207, 620–621. (13) Frauenfelder, H.; Alberding, N. A.; Ansari, A.; Brauenstein, D.; Cowen, B. R.; Hong, M. K.; Iben, I. E. T.; Johnson, J. B.; Luck, S.; Marden, M. C.; Mourant, J. R.; Ormos, P.; Reinisch, L.; Scholl, R.; Schulte, A.; Shyamsunder, E.; Sorensen, L. B.; Steinbach, P. J.; Xie, A.; Young, R. D.; Yue, K. T. J. Phys. Chem. 1990, 94, 1024–1037. (14) Newman, R. C.; Kauzmann, W.; Zipp, A. J. Phys. Chem. 1973, 77, 2687–2691. (15) Henry, E. R.; Sommer, J. H.; Hofrichter, J.; Eaton, W. A. J. Mol. Biol. 1983, 166, 443–451. (16) Marden, M. C.; Hui Bon Hoa, G.; Stetzkowski-Marden, F. Biophys. J. 1986, 49, 619–627. (17) Hamdane, D.; Kiger, L.; Dewilde, S.; Green, B. N.; Pesce, A.; Uzan, J.; Burmester, T.; Hankeln, T.; Bolognesi, M.; Moens, L.; Marden, M. C. J. Biol. Chem. 2003, 278, 51713–51721. (18) Hamdane, D.; Vasseur-Godbillon, C.; Baudin-Creuza, V.; Hui Bon Hoa, G.; Marden, M. C. J. Biol. Chem. 2007, 82, 6398–6404. (19) Kleinert, T.; Doster, W.; Leyser, H.; Petry, W.; Schwarz, V.; Settles, M. Biochemistry 1998, 37, 717–733. (20) Nienhaus, K.; Deng, P.; Kriegl, J. M.; Nienhaus, G. U. Biochemistry 2003, 42, 9647–9658.
Corresponding Author
*Phone: (33)-1-49-59-56-63. E-mail:
[email protected]. Corresponding author address: Inserm, U779, Hopital de Bicetre, 78 rue du General Leclerc, 94275 Le Kremlin-Bic^etre, France. Present Addresses
† Laboratoire d’Enzymologie et Biochimie Structurales, CNRS, 1 avenue de la Terrasse, 91198 Gif-sur-Yvette, France.
’ ACKNOWLEDGMENT This work was supported by Inserm (Institut National de la Sante et de la Recherche Medicale) and the University of Paris11. ’ ABBREVIATIONS: Mb: myoglobin ’ REFERENCES (1) Austin, R. H.; Beeson, K. W.; Eisenstein, L.; Frauenfelder, H.; Gunsalus, I. C. Biochemistry 1975, 14, 5355–5373. 3923
dx.doi.org/10.1021/jp107168b |J. Phys. Chem. B 2011, 115, 3919–3923
